“When you can measure what you are talking about
and express it in numbers, you know something about it”.
Lord Kelvin, 1824-1907
The quote from Lord Kelvin couldn’t be more accurate today in the time of social media and with the increasing volume of user created content on the web, because everything that we as web users do online can be measured and used to map our behaviour, opinions and attitudes, and actions, to know something and everything about us.
Everything that we do online is stored and analyzed. Web giants like Google and Facebook know more about us than our relatives do. Everytime we do a web search, we tell Google what we are interested in or what we are afraid of (because that’s the type of queries that we do). Everytime we write a status update or “like” something on the web, we’re telling Facebook what we are doing and what we like. Google and Facebook use this information as a commodity, to improve the service that they are selling to advertisers. So if you thought that using Google and Facebook was free, you couldn’t be more wrong. Someone has said that if you don’t pay for a service or a product, it usually means that you are the product. The more we tell about ourselves and about our interests, the better Google and Facebook know who we are, and the better product Google and Facebook has. I’m not saying that there’s anything wrong with that, as long as everyone knows what they are signing up for.
The content that we as web users are constantly creating online is not just a product that web giants can use. Some of the content can be analyzed by anyone with the right tools and right knowledge, and this opens up new exciting opportunities for researchers and companies to map our opinions, attitudes, and behaviour on the web. We can analyze for instance how middle aged men in Finland are doing by using http://www.wefeelfine.org/. This site searches for blog entries and automatically seeks for words describing feelings. Obviously this kind of analysis is not always accurate, as some words can have multiple meanings, but still We feel fine is an exciting first step using the technology available today.
Using the API that some social media sites provide we can tap into the constant stream of messages that are being published, we can download the data, and analyze it. We could also just buy the data from data providers like Gnip (http://gnip.com/), but we would need to have really deep pockets for that, and I mean really deep. One of the most exciting services that also provide a fairly generous API for data collection is Twitter. Because of Twitter’s API a multitude of services have been built using data from Twitter and by combining it with some other data. Have a look at http://trendsmap.com/ or http://tweereal.com/ to see some mashup with data from Twitter.
By collecting all tweets that mention a certain brand we can analyze what people are saying about that brand and by mapping the network of the tweets we can find who the opinion leaders are. Have a look at this blog entry (only in Swedish) for an example of these type of analyses. Both providing valuable information for any organization. But perhaps some of the most exciting services are those that do sentiment analysis using Twitter data. One of the first services of this kind was and still is: http://www.sentiment140.com/search?hl=en&query=finland. Using Finland as an example here you can see whether people that have mentioned Finland in a tweet have mentioned Finland in a positive or negative context. Other examples include: http://www.csc.ncsu.edu/faculty/healey/tweet_viz/tweet_app/ and http://smm.streamcrab.com/results?search=finland&pooling=1. I’ve used Finland as an example in these queries, but just as easily we could map what people think about some universities, brands, companies, celebrities, etc.
Play around with the services mentioned here. You may be surprised by the amount of data that is out there, just waiting to be analyzed.
/Kim, @kholmber